Download presentation
Presentation is loading. Please wait.
Published byPatience Lang Modified over 8 years ago
1
NAWQA Program U.S. Department of the Interior U.S. Geological Survey Regression models for explaining and predicting concentrations of organochlorine pesticides in whole fish from United States streams Lisa Nowell, Charles Crawford, Naomi Nakagaki, Gail Thelin, and David Wolock
2
NAWQA Program Today’s talk Organochlorine pesticides and the NAWQA program Regression models Model applications: dieldrin as example
3
NAWQA Program OC pesticides in NAWQA Organochlorine (OC) pesticides –Most uses cancelled during 1972-1988 NAWQA sampling for OC pesticides –Targeted studies in relation to land use –Whole fish and bed sediment
4
NAWQA Program Occurrence in whole fish AgriculturalUrbanUndeveloped
5
NAWQA Program Declining DDT in whole fish 1969 500 0 1000 197419791984198919941999 Concentration (µg/kg) Median DDT Wildlife Benchmark
6
NAWQA Program Potential effects on wildlife One or more benchmarks exceeded No benchmarks exceeded
7
NAWQA Program Summary OCs in whole fish have declined nationally Low levels are still prevalent in streams Potential for adverse effects Implications Need for national assessment Predictive models may be an alternative to widespread direct monitoring
8
NAWQA Program Today’s talk Organochlorine pesticides and the NAWQA program Regression models Model applications: dieldrin as example
9
NAWQA Program Objectives Extend targeted NAWQA studies into a comprehensive national assessment –Explain the distribution of OCs in whole fish –Predict OC concentrations in whole fish in unmonitored streams Approach Develop statistical regression models –p,p’-DDX, p,p’-DDE, p,p’-DDT –cis-chlordane, trans-nonachlor –dieldrin
10
NAWQA Program Model development data: dieldrin Whole fish Multiple taxa 650 streams in 43 Study Units
11
NAWQA Program Explanatory variables Elapsed time since 1966 Fish lipid content Fish taxon Source variables ** Watershed variables ** -Physical -Hydrologic -Soil -Climate -Agricultural management practices Regional variables ** ** Nationwide data sources
12
NAWQA Program Past agricultural use intensity: aldrin + dieldrin
13
NAWQA Program Model Results Pesticide compound Variability explained Most important variables p,p’-DDX 47% Lipid, agricultural use, population density p,p’-DDE 51% p,p’-DDT 41% cis-Chlordane 64% Lipid, termite-urban, population density trans-Nonachlor 64% Lipid, population density Dieldrin 64% Lipid, agricultural use, forested land
14
NAWQA Program Predicted dieldrin in whole fish (6.2% fish lipid) Predicted dieldrin concentration (µg/kg) < 5 5 - 25 >25 - 120 >120
15
NAWQA Program Effect of lipid content on predictions 6.2% lipid 4.2% lipid
16
NAWQA Program Today’s talk Organochlorine pesticides and the NAWQA program Regression models Model applications: dieldrin as example
17
NAWQA Program Dieldrin in whole fish at 6.2% lipid: probability of exceeding 120 µg/kg <5% 5-25% >25-50% >50-75% >75% Probability
18
NAWQA Program Dieldrin in whole fish at 4.2% lipid: probability of exceeding 25 µg/kg 5-25% >75% <5% >50-75% >25-50% Probability
19
NAWQA Program Conclusions Most models explain 50-65% of variability Explanatory variables include lipid, past agricultural use, urban use surrogates Limitations of the models Predictive regression models may be a useful tool in planning future monitoring
20
NAWQA Program For more information: NAWQA pesticide national synthesis project: http://ca.water.usgs.gov/pnsp/ Dieldrin fish model: http://pubs.usgs.gov/sir/2006/5020/ Dr. Lisa Nowell, Chemist U.S. Geological Survey Placer Hall, 6000 J Street Sacramento, CA 95819 (916) 278-3096 lhnowell@usgs.gov
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.